/// <summary>
/// Find value within text block
/// </summary>
/// <param name="textBlock">Text block as string</param>
/// <param name="keys">List of keys to search for</param>
/// <param name="valueWhiteList">List of allowed values</param>
/// <param name="charsToRemove">Chars to remove from a key while comparing</param>
/// <returns>Result instance, which fits the most</returns>
public static ExtractionResult FindInLine(string textBlock, IList <ExtractionKey> keys, IList <ExtractionValue> valueWhiteList, Func <string, bool> validateValue = null, string charsToRemove = ".:;", bool forceWhiteList = false, int minResultLenght = 3)
{
var values = new List <ExtractionResult>();
// Clean keys
foreach (var charToRemove in charsToRemove)
{
foreach (var key in keys)
{
key.Key = key.Key.Replace(charToRemove.ToString(), "");
}
}
var lines = textBlock.Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries);
foreach (var line in lines)
{
// Split words
var words = line.Split(new char[] { ' ' }, StringSplitOptions.RemoveEmptyEntries);
// Go though all words, SKIP the last one
for (int i = 0; i < words.Length - 1; i++)
{
var word = words[i];
var cleanedWord = word;
var keyMatched = false;
var similarity = 0;
ExtractionKey matchedKey = null;
foreach (var charToRemove in charsToRemove)
{
cleanedWord = cleanedWord.Replace(charToRemove.ToString(), "");
}
foreach (var key in keys)
{
// Compare word an key
similarity = LevenshteinDistance.Compute(key.Key, cleanedWord);
if (similarity <= 3)
{
keyMatched = true;
matchedKey = key;
break;
}
}
if (keyMatched)
{
// Continue line loop
for (i = i + 1; i < words.Length; i++)
{
var valueString = words[i];
if (valueString.Length < minResultLenght)
{
continue;
}
// Check value-string is not similar to a key
if (valueString != null && valueString.Length >= 3)
{
foreach (var key in keys)
{
if (key.Key.Length <= 3)
{
continue;
}
// Compare value and keys
var keyValueDistance = LevenshteinDistance.Compute(key.Key, valueString);
if (keyValueDistance <= 3)
{
continue;
}
}
}
// Validate values
if (validateValue == null || validateValue(valueString))
{
if (forceWhiteList && !valueWhiteList.Any(x => x.Value == valueString))
{
continue;
}
// Add as validated value
if (validateValue != null)
{
valueWhiteList.Add(new ExtractionValue {
Value = valueString
});
}
var value = new ExtractionResult
{
KeyDistance = similarity,
Key = matchedKey,
Value = valueWhiteList.FirstOrDefault(x => x.Value == valueString),
OriginalValue = valueString,
CleanedKey = cleanedWord
};
value.ValueMatched = value.Value != null;
values.Add(value);
// Break loop, because a word was found
i = int.MaxValue - 2;
break;
}
}
// Reset matches
keyMatched = false;
similarity = 0;
matchedKey = null;
}
}
}
// Return most similar value
return(values.OrderByDescending(x => x.ValueMatched).ThenBy(x => x.KeyDistance).FirstOrDefault());
}
/// <summary>
/// Find value within text block and retrieve it from the next line
/// </summary>
/// <param name="textBlock">Text block as string</param>
/// <param name="keys">List of keys to search for</param>
/// <param name="valueWhiteList">List of allowed values</param>
/// <param name="charsToRemove">Chars to remove from a key while comparing</param>
/// <returns>Result instance, which fits the most</returns>
public static ExtractionResult FindInNextLine(string textBlock, IList <ExtractionKey> keys, IList <ExtractionValue> valueWhiteList, Func <string, bool> validateValue = null, string charsToRemove = ".:;", int minSplitChars = 3, bool forceWhiteList = false)
{
var values = new List <ExtractionResult>();
// Clean keys
foreach (var charToRemove in charsToRemove)
{
foreach (var key in keys)
{
key.Key = key.Key.Replace(charToRemove.ToString(), "");
}
}
var lines = textBlock.Split(new char[] { '\r', '\n' }, StringSplitOptions.RemoveEmptyEntries);
var splittedReversedLines = new List <IList <string> >();
foreach (var line in lines)
{
splittedReversedLines.Add(SplitLine(line, ' ', minSplitChars, 2).Reverse().ToList());
}
var lineIndex = 0;
ExtractionKey matchedKey = null;
int distance = 0;
var matches = new List <FindNextLineMatch>();
foreach (var line in splittedReversedLines)
{
int wordIndex = 0;
foreach (var word in line)
{
var cleanedWord = word;
foreach (var charToRemove in charsToRemove)
{
cleanedWord = cleanedWord.Replace(charToRemove.ToString(), "");
}
foreach (var key in keys)
{
distance = LevenshteinDistance.Compute(key.Key, cleanedWord);
if (distance <= 3)
{
matchedKey = key;
break;
}
}
if (matchedKey != null)
{
matches.Add(new FindNextLineMatch
{
Key = matchedKey,
LineIndex = lineIndex,
WordIndex = wordIndex,
OriginalKey = cleanedWord,
Distance = distance
});
matchedKey = null;
break;
}
wordIndex++;
}
lineIndex++;
}
foreach (var match in matches)
{
if (splittedReversedLines.Count > match.LineIndex + 1)
{
var line = splittedReversedLines[match.LineIndex + 1];
if (line.Count > match.WordIndex)
{
var valueString = line[match.WordIndex];
if (validateValue == null || validateValue(valueString))
{
if (forceWhiteList && !valueWhiteList.Any(x => x.Value == valueString))
{
// Skip not validated match
}
else
{
// Add as validated value
if (validateValue != null)
{
valueWhiteList.Add(new ExtractionValue {
Value = valueString
});
}
var value = new ExtractionResult
{
KeyDistance = match.Distance,
Key = matchedKey,
Value = valueWhiteList.FirstOrDefault(x => x.Value == valueString),
OriginalValue = valueString,
CleanedKey = match.OriginalKey
};
value.ValueMatched = value.Value != null;
values.Add(value);
}
}
}
}
}
// Return most similar value
return(values.OrderByDescending(x => x.ValueMatched).ThenBy(x => x.KeyDistance).FirstOrDefault());
}
